Large-scale Medical Image Analysis

The main objective of the project is to develop a large–scale analysis framework enabling transferring 2D visual data analysis techniques to very large scales but also to 3D, 4D and multimodal imaging with high efficiency, and by increasing the scalability of existing tools and algorithms to large amounts of data. The goal is to reduce the execution time of image analysis tasks and enhance interactions with end–users by selecting the right underlying infrastructures such as centralized servers, distributed clusters or even cloud computing. Bandwidth, storage capacity, processor power and main memory may all have their limiting effects and choosing an infrastructure should be facilitated with the results of LaMIA.

Although the developed techniques have an application focus on medical image analysis and retrieval, they are expected to provide tools for managing visual information in a large variety of domains from fundamental research to industrial applications.

The developed techniques will be promoted by being transferred to several research projects on medical image analysis and medical information retrieval. Resulting methods will serve as a basis for new project proposals on large–scale medical information management on a national as well as international level.

Figure 1. Hadoop Map/Reduce


Henning Müller

Adrien Depeursinge

Ivan Eggel

Dimitrios Markonis


University Hospitals of Geneva
  • Pierre-Alexandre Poletti, Emergency Radiology
  • Antoine Geissbuhler, Medical Informatics